Fig. 4

Short DNA sequence patterns enriched in broadly active enhancers have high GC content. a The log₂ of the fold enrichment for each 6-mer (N = 4096) is plotted against its GC-content for comparison of broadly active vs. non-GC-matched background. b The same as (a) but comparing broadly active enhancers to GC-matched background regions. c Mean enhancer GC content is significantly correlated with breadth of activity among enhancers (Spearman’s ρ = 0.25, P < 2.2E–308). Enhancers were assigned to bins based on the number of contexts in which they were active (x-axis); each bin is labeled with the highest activity present in that bin, and the number of regions in it. d Same as (a), but comparing broadly active to context-specific enhancers. The number of 6-mers in each GC content bin is: 64 for 0 and 1, 384 for 0.167 and 0.833, 960 for 0.333 and 0.666, and 1280 for 0.5